DE102005039184B4 - Method for evaluating a cinematographic image series of the heart, magnetic resonance tomography apparatus and computer program - Google Patents

Abstract

A method of evaluating a cinematographic image series (1) of the heart (3), wherein the series of images is a short-axis incision series of the heart, during a cardiac cycle, comprising the following steps:
Recording the image series (1) with a recording unit in which image series (1) healthy heart muscle tissue (24, 44) is distinguishable from damaged myocardial tissue (25, 45) due to contrast agent-free and contrast-medium-marked regions,
Processing of at least two individual images (5, 5 ') of the image series (1) by means of a data processing unit, in each of the individual images (5, 5') in the region of the heart muscle, in the damaged heart muscle tissue (25, 45) and healthy myocardial tissue ( 24, 44), by means of a pattern recognition algorithm (6) at least a region of the healthy myocardial tissue (24, 44) is detected by defining a radius (29, 29 ') and along the radius (29, 29') measuring a radial diameter (31, 31 ', 51, 51') of the area of the healthy cardiac muscle tissue,
- Detecting a dynamic change (7) of the radial diameter (31, 31 ', 51, ...

Description

The invention relates to a method for evaluating a cinematographic image series of the heart, in which healthy myocardial tissue is shown distinguishable from damaged, in particular in a cine-delayed enhancement MRI image series of the heart. Furthermore, the invention relates to a magnetic resonance tomography apparatus and a computer program for carrying out such a method.

In the closest publication Proc. Intl. Soc. Mag. Reson. Med. 13, p. 236 (2005) describes a method in which a cinematographic or cine series of images of the heart is recorded 10 to 20 minutes after an intravenous administration of gadopentate dimeglumine with a TrueFISP sequence. This image series is called Cine-Delayed Enhancement series. This results in a series of images of the heart during a cardiac cycle in which ischemia-damaged myocardial tissue is characterized by a contrast enhancement. The evaluation of such a series of images is done manually so far by the user, which is associated with a high staff and time. The extent of the scar tissue and the contractility of the entire heart wall are quantified.

The DE 102 56 208 A1 discloses a method for the velocity-resolved flow measurement during a motion cycle in magnetic resonance tomography, in which an anatomical image series of a selected region and, at the same time, a velocity-resolved image series of this region are measured. In a second step, both series of images are superimposed in such a way that each speed-resolved image is integrated in the temporally corresponding anatomical image.

The WO 01/20305 A1 discloses a special method of optical tomography, in which an object to be examined is irradiated with optical waves near the IR region and the optical waves scattered and exiting from the object under examination are measured. From this, sectional images or temporal image series of the object to be examined can be made.

The US 5,803,914 A discloses a system for data acquisition, data processing, and rendering of so-called gated SPECT image data used to diagnose coronary artery diseases in nuclear medicine. Two parameters are used to evaluate a possible coronary heart disease, namely information that relates to the perfusion of the myocardium, as well as information related to the movement of the myocardial wall.

It is the object of the present invention to provide a method with which a series of images of the heart during a cardiac cycle in which healthy myocardial tissue is shown distinguishable from damaged, is evaluated to essential parts automatically and the user medically relevant information on the degree of damage are output ,

This object is achieved according to the invention by the features of the independent claims. The dependent claims further form the central idea of the invention in a particularly advantageous manner.

• – Aufnehmen der Bildserie mit einer Aufnahmeeinheit, wobei sich bei der Bildserie gesundes Herzmuskelgewebe von geschädigtem Herzmuskelgewebe unterscheidbar aufgrund von kontrastmittelfreien und kontrastmittelmarkierten Bereichen darstellt,
• – Verarbeiten von zumindest zwei Einzelbildern der Bildserie mithilfe einer Datenverarbeitungseinheit, indem an jedem der Einzelbilder mittels eines Mustererkennungsalgorithmus zumindest ein Bereich des gesunden Herzmuskels und/oder des geschädigten Herzmuskels erfasst wird, und indem zumindest eine beschreibende Größe gemessen wird, die eine geometrische Eigenschaft des Bereiches beschreibt, der durch den Mustererkennungsalgorithmus erfasst worden ist,
• – Erfassen einer dynamischen Änderung der beschreibenden Größe, und
• – Ausgeben einer medizinisch relevanten Information, die aus der dynamischen Änderung der beschreibenden Größe gewonnen wird.

According to the invention, a method for evaluating a cinematographic image series of the heart during a heart cycle is claimed, comprising the following steps:

• Recording the series of images with a recording unit, wherein in the series of images healthy myocardial tissue is distinguishable from damaged myocardial tissue due to contrast-agent-free and contrast-medium-marked areas,
• Processing at least two individual images of the image series with the aid of a data processing unit by detecting at least one region of the healthy heart muscle and / or the damaged heart muscle on each of the individual images by means of a pattern recognition algorithm, and by measuring at least one descriptive variable representing a geometric property of the region described by the pattern recognition algorithm,
• - Detecting a dynamic change of descriptive size, and
• - Issuing medically relevant information, which is obtained from the dynamic change of the descriptive size.

As a result of this method, significant and labor-intensive steps that previously required a manual evaluation are solved semi-automatically or automatically during the evaluation of the image series, so that time and costs are saved. In this case, the descriptive variable characterizes geometric information of the region recognized by the pattern recognition algorithm and can be, for example, a diameter, the width or the area of the region. The pattern recognition algorithm can work automatically or semi-automatically. For example, if it is designed as an automatic segmentation algorithm, it can recognize and optionally mark the well distinguishable contrast-free and contrast-marked areas of the heart muscle. In a interactive design of the segmentation algorithm, the user himself can set points that serve as starting points for the segmentation algorithm, and optionally intervene manually and make corrections. The pattern recognition algorithm does not necessarily have to be designed as a segmentation algorithm. By way of example, it may be sufficient for the pattern recognition algorithm to mark the start and the end of contrast-marked or -free regions along predetermined lines-for example along cutting lines-and mark them if appropriate. This makes it possible to determine the diameter of areas.

Since the evaluation takes place on at least two images of the image series, the change in the descriptive variable can be tracked over the cardiac cycle. From the dynamic change of size medically relevant information can be obtained whose information content goes beyond what would result from the evaluation of a single static image. Subsequently, the information is output to the user.

In a preferred embodiment, the medically relevant information is a graphical and / or a tabular representation of the dynamic change of the descriptive variable. This allows the user to see quantitatively at a glance how the descriptive size changes over the time of the cardiac cycle.

Advantageously, as a medically relevant information, a value is output from the dynamic change of the descriptive variable that belongs to a particular point in time. The time may be a particularly characteristic time during the heart cycle, such as the time of the end systole or end diastole.

In a further embodiment of the invention, the medically relevant information is the percentage change in size between two specific times, such as the end systole and the end diastole. This allows the user to be easily informed of the change in the particular size during the cardiac cycle.

In a further embodiment of the invention, the medically relevant information is a maximum or a minimum value from the dynamic change of the descriptive variable or the percentage change of the descriptive variable with respect to the maximum and minimum value. Again, the user is informed in a simple and concise manner about the change of the particular size, for example about the radial diameter of the infarcted scar, during the cardiac cycle.

In a preferred embodiment, the cinematographic image series is a short-axis section series of the heart. Depending on the clinical question, however, it is also possible to select a series of images of another sectional plane, for example a longitudinal-axis section or a four-chamber section series.

In the short-axis section series, a radius is advantageously set in a certain direction, and a descriptive quantity is a radial diameter of the healthy heart muscle along the radius. This advantageously determines the percentage change in the heart muscle thickness of the healthy heart muscle during a cardiac cycle. From this information, the user can see what contractility the healthy heart muscle has in the direction of the radius. Especially in areas of the heart muscle where both healthy and damaged myocardial tissue is present, this information provides information on the residual contractility and vitality of healthy heart muscle tissue and may play a crucial role in further treatment planning.

In a further advantageous embodiment of the invention, a plurality of radii are set in different directions, and for each of these radii a directional percentage change in the heart muscle thickness of the healthy heart muscle is detected. This provides information about the spatial distribution of contractility and contractility reductions. Preferably, the directional percent change in cardiac muscle thickness is displayed in a bulls-eye plot to provide the user with a graphically concise representation of the medically important information. The bulls-eye plot can be color-coded.

In a further embodiment, as in the measurement of the healthy heart muscle, a radial diameter of the damaged heart muscle is measured along a radius in a specific direction. Advantageously, the percentage change in cardiac muscle thickness of the damaged myocardium along this radius is determined during a cardiac cycle. This provides the user with inferences about any residual contractility that may still exist or about a possibly threatening heart wall rupture. Here, too, it is preferable to define a plurality of radii in different directions, along which the percentage change in the heart muscle thickness of the damaged heart muscle is detected. This allows information about the spatial extent of the damaged heart muscle tissue win. The representation preferably takes place in a bulls-eye plot, which can be color-coded.

In a preferred embodiment, the cinematographic image series of the heart is a Cine-Delayed Enhancement MRI series, since in this series, healthy and damaged myocardial tissue is well distinguishable. However, other series of pictures of the heart are also suitable as long as healthy and damaged myocardial tissue is distinguishable. This can also be ensured, for example, with a gated-SPECT-CT display of the heart.

Further, according to the present invention, an MRI apparatus suitable for carrying out the method according to the above claims is claimed.

Also claimed is a computer software product implementing a method according to the above claims when running on a computing device connected to this magnetic resonance tomography device.

Further advantages, features and characteristics of the present invention will now be explained in more detail by means of exemplary embodiments with reference to the accompanying drawings. Show it:

1 schematically the procedure,

2 schematically an image of a short-axis incision series of the heart at the time of the final systole,

3 schematically a picture of a short-axis incision series of the heart at the time of end diastole, and

4 a bulls-eye plot to indicate the contractility of the healthy heart muscle.

1 schematically shows the procedure, as in the evaluation of a cinematographic image series 1 of the heart 3 during a cardiac cycle. The starting point is a series of pictures 1 of the heart 3 during a cardiac cycle. Preferably, the image series 1 a Cine Delayed Enhancement MRI series. In this series 1 Damaged cardiac muscle tissue significantly enhances contrast agents, creating a strong contrast between damaged and healthy myocardial tissue, resulting in a pattern recognition algorithm 6 can reliably differentiate between areas of damaged and healthy myocardial tissue.

The pattern recognition algorithm 6 is preferably on all frames of the series 5 . 5 ' applied. This allows the areas of healthy and / or damaged myocardium to be detected, and their changes throughout the cardiac cycle can be tracked. In practice, it is usually sufficient if, instead of changing the areas, the change of a certain size is followed, which describes the geometry of the areas. For example, the change in a radial diameter of healthy myocardial tissue during a cardiac cycle has a high significance for the contractility and thus also for the vitality of the healthy myocardial tissue. Therefore, after the frames 5 . 5 ' with the pattern recognition algorithm 6 have been processed, the dynamic change 7 the determined size, as described in detail with reference to the following figures.

It is often advantageous to determine several quantities that describe geometric properties of different areas in the heart muscle. For example, at one location of the heart, the inner layer of the heart muscle may contain damaged myocardial tissue while the outer layer of the heart muscle is still intact. In this case, it is advantageous to determine two quantities, one describing the diameter of the healthy heart muscle and the other describing the diameter of the damaged heart muscle. These two quantities can be related to each other, for example, to determine their percentage of the total diameter of the heart muscle.

In a further step, the user receives medically relevant information resulting from the dynamic change 7 the size yields, spent. This can be, for example, a graphical representation 9 the change of the particular size itself or a tabular representation 11 be. It is advantageous if values 15 , which belong to characteristic times of the heart cycle, must be marked separately. From the dynamic change 7 can also percentages and / or absolute values 13 determine the change in the particular size during a cardiac cycle.

2 schematically shows a short-axis section of the heart at the time of the end systole. The heart muscle of the left ventricle 21 and the right ventricle 23 are contracted. The damaged heart muscle tissue 25 is located on the inner layer of the heart and is made of healthy myocardial tissue 24 surround. In this particular example, the pattern recognition algorithm pulls from a central point 27 in the left ventricle 21 from two radii 29 and 29 ' in two different directions. Along these two radii 29 and 29 ' The pattern recognition algorithm captures the areas of the healthy and damaged heart muscle, respectively, whose radial diameters 31 . 31 ' respectively. 33 . 33 ' can then be measured.

In 3 is shown schematically another short-axis section of the heart from the same cinematographic series of images of the heart, from the well 2 taken at the time of Enddiastole. The heart muscle of the left 21 and the right ventricle 23 is relaxed. Also in this picture, the areas of the healthy and the damaged heart muscle are along the same radii 29 and 29 ' from the same starting point 27 go out, recorded. The radial diameter 51 . 51 ' denote the thickness of healthy myocardial tissue, the radial diameter 53 . 53 ' the thickness of the damaged heart muscle tissue.

While along the radius 29 the radial diameter of the healthy heart muscle 31 and 51 change significantly between end-systole and end-diastole, remain the radial diameter 31 ' and 51 ' along the radius 29 ' largely the same. This means that in this particular example, the contractility of the myocardial tissue along the radius 29 ' is clearly limited. This information may play a crucial role in further treatment planning, such as recanalization of the corresponding coronary artery.

4 shows you one 2 and 3 appropriate bulls-eye plot in which the contractility of the healthy heart muscle is encoded in various radial directions. This is one possible way of presenting the results, which allows the user to locate function limitations of the heart muscle in a particularly simple manner. In the bulls-eye plot shown here, the heart muscle has been divided into eight octants. In the area 60 is the contractility of healthy myocardial tissue - matching the location of the radius 29 ' - severely restricted while in the fields 64 the heart muscle tissue contracted normally. In the intermediate areas 62 - matching the radius 29 and 63 Although the contractility of the heart muscle is limited, but not to the extent as in the area 60 ,

Claims (21)

Method for evaluating a cinematographic image series ( 1 ) of the heart ( 3 ), wherein the series of images is a short-axis incision series of the heart, during a heart cycle, comprising the following steps: - taking the image series ( 1 ) with a recording unit, in which image series ( 1 ) healthy heart muscle tissue ( 24 . 44 ) of damaged myocardial tissue ( 25 . 45 ) is distinguishable on the basis of contrast-free and contrast-marked regions, - processing of at least two individual images ( 5 . 5 ' ) of the image series ( 1 ) by means of a data processing unit, by applying to each of the individual images ( 5 . 5 ' ) in the region of the heart muscle, in the damaged heart muscle tissue ( 25 . 45 ) and healthy myocardial tissue ( 24 . 44 ) by means of a pattern recognition algorithm ( 6 ) at least a portion of the healthy myocardial tissue ( 24 . 44 ) is detected by a radius ( 29 . 29 ' ) and along the radius ( 29 . 29 ' ) a radial diameter ( 31 . 31 ' . 51 . 51 ' ) of the area of the healthy myocardial tissue, - detecting a dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ), and - issuing medically relevant information resulting from the dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ) is won. The method of claim 1, wherein the medically relevant information is a graphical ( 9 ) and / or a tabular representation ( 11 ) of the dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ). The method of claim 1, wherein the medically relevant information is a value ( 15 ) from the dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ) at a certain point in the dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ) belongs. The method of claim 1, wherein the medically relevant information is the percentage change ( 13 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ) between two specific times. The method of claim 4, wherein the two particular times are the times of the end systole and the end diastole. The method of claim 1, wherein the medically relevant information is a maximum or a minimum value from the dynamic change ( 7 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ). The method of claim 1, wherein the medically relevant information is the percentage change ( 13 ) of the radial diameter ( 31 . 31 ' . 51 . 51 ' ) with respect to a maximum and a minimum value from the dynamic change ( 7 ). Method according to one of Claims 1 to 7, in which, in the short-axis section series, a plurality of radii ( 29 . 29 ' ) are set in different directions, and at each of these radii ( 29 . 29 ' ) each have a radial diameter ( 31 . 31 ' . 51 . 51 ' ) of healthy heart muscle tissue ( 24 . 44 ) and an associated directional percent change in cardiac muscle thickness of healthy myocardial tissue ( 24 . 44 ) is detected. The method of claim 8, wherein the directional percentage changes in cardiac muscle thickness of the healthy myocardial tissue ( 24 . 44 ) in a bulls-eye plot. Method according to one of claims 1 to 9, comprising the following further steps: - Processing of at least two individual images ( 5 . 5 ' ) of the image series ( 1 ) using the data processing unit, by adding to each of the individual images ( 5 . 5 ' ) by means of the pattern recognition algorithm ( 6 ) at least one further area of the damaged heart muscle tissue ( 25 . 45 ) is detected by moving along the radius ( 29 . 29 ' ) another radial diameter ( 33 . 33 ' . 53 . 53 ' ) of the further area of the damaged heart muscle tissue ( 25 . 45 ), - detecting a dynamic change ( 7 ) of further radial diameter ( 33 . 33 ' . 53 . 53 ' ), and - output of further medically relevant information resulting from the dynamic change ( 7 ) of further radial diameter ( 33 . 33 ' . 53 . 53 ' ) is won. Method according to Claim 10, in which the further medically relevant information is a graphical (9) and / or a tabular representation ( 11 ) of the dynamic change ( 7 ) of further radial diameter ( 33 . 33 ' . 53 . 53 ' ). Method according to one of Claims 10, in which the further medically relevant information has a value ( 15 ) from the dynamic change ( 7 ) of the further radial diameter ( 33 . 33 ' . 53 . 53 ' ) at a certain point in the dynamic change ( 7 ) of the further radial diameter ( 33 . 33 ' . 53 . 53 ' ) belongs. The method of claim 10, wherein the further medically relevant information is the percentage change ( 13 ) of the further radial diameter ( 33 . 33 ' . 53 . 53 ' ) between two specific times. The method of claim 13, wherein the two particular times are the times of the end systole and the end diastole. Method according to Claim 10, in which the further medically relevant information is a maximum or a minimum value from the dynamic change ( 7 ) of the further radial diameter ( 33 . 33 ' . 53 . 53 ' ). The method of claim 10, wherein the further medically relevant information is the percentage change ( 13 ) of the descriptive quantity with respect to a maximum and a minimum value from the dynamic change ( 7 ) of the further radial diameter ( 33 . 33 ' . 53 . 53 ' ). Method according to one of claims 10 to 16, wherein in the short-axis section series a plurality of radii ( 29 . 29 ' ) are set in different directions, and at each of these radii ( 29 . 29 ' ) each have a further radial diameter ( 33 . 33 ' . 53 . 53 ' ) of the damaged heart muscle tissue ( 25 . 45 ) and detecting an associated directional percent change in cardiac muscle thickness of the damaged myocardial tissue. The method of claim 17, wherein the directional percentage changes in heart muscle thickness of the damaged myocardium are displayed in a bulls-eye plot. The method of any one of claims 1 to 18, wherein the short-axis incision series of the heart is a cine-late enhancement MRI series. Magnetic resonance imaging apparatus, which is designed to carry out the method according to one of the above claims 1 to 19 with a data processing unit and with a means for displaying. Computer program for implementing the method according to one of the above claims 1 to 19, when it runs on the data processing unit connected to the magnetic resonance tomography apparatus according to claim 20.

Images